We will demonstrate and evaluate computer-based, interactive knowledge centers as a means of making current medical knowledge conveniently and effectively available to medical professionals. Starting with time-shared systems which reflect the current state of the computer-technology art, we will proceed to investigate their enhancement with the impending technologies of high-capacity "personal" computers controlling knowledge bases stored on videodisks. As a representative area of rapidly developing knowledge we have selected the area of lithium therapy, in which area our prototype time-shared system will make available to practitioners in a clinical setting synopses and possible therapy schedules relevant to problems raised by particular patients. For the initial system, if the users want primary sources, these will be provided in printed form, derived from a very complete Lithium Library. Various expeiments will be conducted evaluating this knowledge dissemination method, compared to more traditional methods dependent on seminars, courses, journal reading and haphazard library use, with respect to measurable physician expertise and respect to specific success in treating patients. Building on this prototype, the research will proceed to enhance and improve it by developing and then applying improved languages, programming environments, and text-structuring methods, specifically designed to facilitate the construction and performance of computer-based, interactive medical knowledge stations. Among other things, we will investigate methods of abstracting graphically represented knowledge and integrating it with text, in a manner which fully exploits the potential of computer-controlled videodisks.